Chapter 10 Computing plant height

10.1 Introduction

Plant height is not very well simulated for the moment. It is computated using the plant width, which depends itself on the plant lai (lai+eai+laisen). Please see Chapter 2 for more details. Figure 10.1 shows the simulated height of wheat in Auzeville, and the variables from which it is computed.

Figure 10.1: Simulated height compared to the observed values, and variables from which it is computed from, in a self-intercrop of Wheat-Wheat

The plant height is important in the case of intercrops because it has high impact on the light interception of both species. Indeed, a sensitivity analysis was performed using the sticRs package on the parameter that links the lai to the height: P_rapforme. The results are shown below (see html version of this book for interactivity):

height_sensi= sticRs::load_sensi(dirpath = "data/rapforme_sensitivity")
if(knitr:::pandoc_to() == "html") {
  gg_objects= 
    lapply(height_sensi$gg_objects, 
           function(x){
             plotly::ggplotly(x, tooltip= "group")
           })
  htmltools::tagList(gg_objects)
}else{
  multiplot(plotlist = height_sensi$gg_objects)
}

As we can see on the above plots, as the plant height is affected, its lai is also impacted afterwards. We can also see that no values of P_rapforme gave satisfactory plant height, so maybe the link between the plant lai and the plant height is not as straightforward as previously thought, and maybe there is another variable that could be better correlated to it.

10.2 Link between height and lai and dry mass

Using the observations from Auzeville, we plotted the evolution of the plant height through time along the lai and the masec (i.e. dry mass) to see if and how they are correlated. The Figure 10.2 shows this evolution with standardized and linearly interpolated observed values for all variables to make it easier to compare.

Figure 10.2: Observed plant height compared to the lai and dry mass for both wheat and pea in sole crop.

The Figure 10.2 shows that the plant height is better correlated to the masec than to the lai. We propose to either use the masec or somcour to compute it rather than the lai.

10.3 Computing plant height

10.3.1 Observations

The plant height could be computed by either using the plant dry mass (masec) or more simply the development stage, derived from cumulated degree days (somcour). The second method could be much simpler to parameterize from field, however, both plants experience the same microclimate in STICS for intercrops, so a plant growing dominated by an another plant could have the same cumulative temperature than a plant in sunlit conditions. Hence, this method would not consider the competition effect in interspecies for height, so it cannot be implemented as long as both plants have the same somcour.

To document further this phenomenon, Figure 10.3 shows that a Pea grown in intercrop with a wheat has a lower height than a Pea grown in solecrop, and that the dry mass is a good proxy to predict this behaviour.

Figure 10.3: Observed Pea height and dry mass for a Pea grown in sole crop or in intercrop with wheat.

Hence, we will implement a computation of the plant height using the plant dry mass (masec).

10.3.2 Stresses

Water and nitrogen stresses can both influence the height of a plant. Figure 10.4 shows that the maximum height is reduced for the wheat grown under nitrogen stress (i.e. no nitrogen feritilization). However, the height and massec relationship are very close for both management, only the dry mass is higher for fertilized wheat, which gives a higher height in the end.

Figure 10.4: Observed relationship between height and dry mass for a wheat grown in sole crop.

Consequently, height is indeed influenced by nitrogen stress, but it appears from our data that it is only through the height~masen_n relationship. Conequently, the height computation shouldn’t be corrected by any nitrogen indices, because this stress is already accounted for in the dry mass computation.

10.3.3 Computation

The new computation is implemented in STICS as a new subroutine (formplante_mas) that is called instead of formplante if the newly defined P_code_shape parameter is equal to 2. This subroutine is largely based on the previous one, conserving all side effects of P_hautmax, P_hautmaxtec, P_largtec and originehaut but add a computation of the plant height using an allometric relationship with the dry mass as follow :
$ hauteur= P_{hautbase} + P_{hautK} masec^{P_{hautA}}$

The P_hautbase parameter is a parameter from formplante, and is used as the intercept of the relationship, i.e the base height when the dry mass is equal to 0. P_hautK is a new parameter that defines the constant of proportionality that reflects the importance of the relationship, and P_hautA the scaling exponent, or the allometric coefficient. Allometric relationships are simple yet powerfull laws used to describe diverse relationships for wide range of situations. Further documentation can be found on the wikipedia webpage.

10.3.4 Results

Two simulations of a wheat in sole crop and a pea in sole crop were run using STICS to asses the potential of the new computation for height modelling, and the outputs are displayed in Figure 10.5.

Figure 10.5: STICS simulation output for a simulation on two sole crops: wheat and pea

The results show that using masec instead of the lai to model plant height gives better results. The model will now be used with this formalism.